VMware Cloud Community
Nitzan_S
Enthusiast
Enthusiast
Jump to solution

Virtual ESXi / Nested Virtualization and EVC

Hi all,

We've tried to install some new virtual ESXi servers in our dev. environment in order to help our R&D.

The problem is that our clusters have to be EVC enabled - Merom Gen. in order for the R&D stuff to work.

I've tried using the following two articles:

http://www.virtuallyghetto.com/2011/07/how-to-enable-support-for-nested-64bit.html

http://www.virtuallyghetto.com/2012/02/how-to-configure-nested-esxi-5-to.html

After I finish the first link, the virtual ESXi is able to run 64-bit VMs. So everything's fine there (Although I did ran into problems with this as well - Apparently you can't have the physical host hosting the virtual ESX running in an EVC in the first place).

So after the virtual ESXi is able to run 64-bit VMs, I'm continuing to the 2nd link and it doesn't work. I'm doing exactly what the guy says there and yet the virtual ESXi is not compatible to the EVC cluster and the 64-bit VM hosted on the virtual ESXi cannot run after the CPUID Mask changes.

Has anyone experienced something similiar? Anyone has any ideas what am I doing wrong or can provide me with some leads on what to do?

Thanks ahead,

Nitzan.

0 Kudos
1 Solution

Accepted Solutions
admin
Immortal
Immortal
Jump to solution

I believe the following option:

cpuid.1.ecx = 00000000000000001010001000011001

should be

cpuid.1.ecx = 00000000000000001010001000111001

View solution in original post

0 Kudos
6 Replies
a_p_
Leadership
Leadership
Jump to solution

Discussion moved from VMware ESXi 5 to Nested Virtualization

0 Kudos
admin
Immortal
Immortal
Jump to solution

Can you post a vmware.log file for one of the ESXi VMs?

0 Kudos
Nitzan_S
Enthusiast
Enthusiast
Jump to solution

I added both the vmware.log of the virtual ESXi (vesx_vmware.log) and of the 64 bit vm hosted on it (64bitvm-vmware.log).

0 Kudos
admin
Immortal
Immortal
Jump to solution

I believe the following option:

cpuid.1.ecx = 00000000000000001010001000011001

should be

cpuid.1.ecx = 00000000000000001010001000111001

0 Kudos
Nitzan_S
Enthusiast
Enthusiast
Jump to solution

You're right. The new option fixed it.

Where did you see that this was the missing option? In the VC Errors it didn't say that this option was the wrong one.

Thanks a lot!

0 Kudos
admin
Immortal
Immortal
Jump to solution

I looked at the source code to see what the requirements for a Merom cluster were.

0 Kudos